Development of a bilingual healthcare chatbot for pregnant women: A comparative study of deep learning models with BiGRU optimization

With the growing demand for healthcare services and a persistent shortage of medical professionals, intelligent systems such as chatbots are gaining relevance in improving patient support. In obstetrics, pregnant women require fast, accessible, and reliable information to monitor their health and the progression of their pregnancy. This study aims to design and evaluate a bilingual chatbot tailored to the healthcare needs of pregnant women, leveraging recent advances in deep learning for natural language processing (NLP). We developed and compared five deep learning architectures—artificial neural networks (ANN), long short-term memory (LSTM), bidirectional LSTM (BiLSTM), gated recurrent units (GRU), and bidirectional GRU (BiGRU)—to identify the most suitable model for chatbot implementation. Each model was trained on a bilingual dataset of pregnancy-related questions and answers, and evaluated using accuracy, computational efficiency, and contextual relevance of responses. The BiGRU model achieved the highest performance, demonstrating superior accuracy and response efficiency over the other models. It consistently delivered context-aware, personalized answers in both languages, showing its robustness in handling sequential healthcare queries. These findings suggest that BiGRU networks offer a promising solution for building intelligent, bilingual healthcare chatbots aimed at supporting pregnant women. Future work will focus on expanding the dataset, incorporating voice-based input, and deploying the chatbot in real-world healthcare settings for clinical validation. • Pregnant women require fast and reliable access to essential information about their pregnancy. • Recent progress in deep learning algorithms has enabled the development of chatbots capable of providing tailored, accurate interactions for various healthcare applications. • The study evaluates five deep learning algorithms—ANN, LSTM, BiLSTM, GRU, and BiGRU—specifically for designing a bilingual chatbot to monitor pregnancy progress. • The BiGRU model outperforms other algorithms in terms of accuracy and efficiency in training data.